Docs/Voice/Overview
Voice

Voice

3 min read

Voice lets you answer and steer agents by talking instead of typing. Dictation runs at roughly 150 words a minute against about 40 for a good typist, so the long instruction you would never bother typing takes ten seconds to say. It works in the Director (the Speak button on the prompt bar), in the Cockpit, and on your phone.

Screenshot coming soonThe prompt bar with the Speak button active while dictating an instruction to an agent
The prompt bar with the Speak button active while dictating an instruction to an agent

The trick is the dictionary

Generic dictation mangles technical speech - repo names, library names, your project's jargon. DevThrottle gives you one custom dictionary: teach it your words once and every transcription respects them. The dictionary is shared everywhere you dictate - desktop, browser, and phone - so "Supabase", "devcontainer", or your product's codename comes out right on all of them.

Tip
Seed the dictionary with the ten words you say most - project names, tool names, teammates - and dictation gets dramatically better in the first minute.

Two ways to run it

Transcription needs a speech model, and you choose how that is powered:

  • Bring your own key. Use your own AI account for transcription. Your key stays on your machine - DevThrottle's cloud never stores it.
  • Hosted transcription. We handle the key - zero setup. It is billed from your prepaid credits at the published rate (rate card), activated by your first top-up.

The trade-offs are on hosted vs bring-your-own key.

The privacy line

Note
Voice is the only feature that sends anything off your machine, and only when you use it - and what leaves is your voice audio, never your code. Agents, terminals, and files stay local either way.

Where to go next